Bill Hooker – a staunch supporter and campaigner for Open Data – has published his first mashup. He queries whether it actually is one – and I tend to agree – but the effect is to bring together different sources of information – Nature’s table of antibody suppliers and Google’s custom search. And whatever it is technically it’s an example of how linking information on the web can dramatically enhance our power.
FWIW My view of mashup is informed by the classic ChicagoCrime where two entirely unconnected resources – the city’s crime reports , and Google maps – are linked by common, agreed, metadata – in this case the geographical coordinates of the crime. By accessing the crime data the coordinates are transmitted to Google maps and displayed; alternatively you can query the maps (inluding areas and routes) and access the crime database to find out what sort of crime happened in that geographical objects.
The requirements are simple:
- common, agreed metadata
- public information
- re-usable without restriction (e.g. CC-BY)
then anyone can create a mashup where they discover two sources with the same metadata.
So could it work for chemistry? Of course – we publish the information with common metadata (chemical identity, standardized dictionaries of properties, perhaps some human metadata – persons, institutions, etc.). The software overhead is very light – a few pages of javascript is probably enough – a nice web interface and so on.
There’s one tiny flaw – where’s the public re-usable information? There should be lots – we scientists publish over 1 million compounds per year – that will make a lovely mashup. But then the publishers copyright it and through various techniques – cutting off subscriptions, sending in lawyers, hiring pitbulls put endless obstacles in our way.
But there are growing chinks: Pubchem, ChEBI, the chemical blogosphere, openness of patents, CrystalEye and COD, Blue Obelisk, e-theses. The growing support from outside chemistry. Some of this I’ll publicize, some we’ll discuss in private e-sessions where we plot our next moves on how we can wrest our rightful data back. A little bit of FUD from our side might be useful 🙂
Hi Peter,
I tend to differentiate natural data recombination and forced data recombination of Web Data using the terms: Meshup and Mashup respectively.
When Data is inside the Web (thanks to the dexterity of RDF) it is meshable. Whereas, when it is simply published on the Web is is mashable.
Mashups don’t always expose their raw data sources (but your example at: http://www.sennoma.net/main/edits/ab.html ) does, which opens up another interesting dynamic i.e. the production of Meshups from Mashups.
Here is a quick example (using an RDF Browser that exploits our “Sponger”):
http://demo.openlinksw.com/DAV/home/demo/Public/Queries/DataWeb/antibodies_search_meshup.wqx
Note: There is much more that can be done with all of this when the data is meshed with DBpedia and other sources of RDF based Open Data 🙂
(1) I agree – and I like the idea of meshup. The mashup – as perhaps you imply – is an evolutionary stage towards the meshup – yo which I am fully committed. I evangelised DBPedia to the maths people yesterday. They, of course, know all about graphs but hadn’t – I think – seen the power of evolving semi-structured information. When the world is one big triple store – and that’s a simple vision – then mashups are irrelevant and meshup is simply a synonym.
The key goal is to create and expose enough high quality tripes today to convince people of the usefulness. Here’s an idea for the DBPedia people – “Query of the day” – inviting the world to create exciting queries in DBP.
I have just started looking at your Sponger. If what I think you have done is true – i.e. you have created a complete triple store from the addresses and terms that Bill gave this morning – then I am very impressed. As with all triple stores it’s sometimes difficult to get started. Do you have a simple SPARQL like DBPedia?
FWIW we are actively working on SKOS in e-theses and making very good progress.
If you mail or post some example on how to use the resource I’ll post them. And I think Bill will be impressed as well.
A little or quite off-topic maybe, but you might be interested in http://www.scai.fraunhofer.de/fileadmin/images/bio/bio-flyer/Projektblatt_CSR_02.pdf